Samenvatting
A framework of scientific methodology for specifying the conditions for testing causal relationships in accidents is provided. For experimental control the test variables must be isolated; however, accidents involve interdependently the driver, vehicle and environment. Therefore, in addition to the basic methodology, an organizational model is necessary to coordinate the interrelated system. The discussion concerns driver, vehicle, and environmental factors, experimental designs and problems, naturalistic versus laboratory research, bias and corrective measures and data sampling.